ClawHub
Back to skill

Security audit

Library of Babel

Security checks across malware telemetry and agentic risk

Overview

This skill is a local Library of Babel demo with a documentation inconsistency about storage, but no evidence of hidden, unsafe, or high-impact behavior.

Before installing, understand that the optional codex feature may save text you add into a local codex.json file despite the headline 'No storage' claim. Avoid adding private text unless you are comfortable with it being stored locally, and note that the published artifact appears incomplete because referenced Python demo files are not included.

SkillSpector

Vulnerability Patterns
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Description-Behavior Mismatch

Medium
Confidence
95% confidence
Finding
The manifest and top-level documentation repeatedly describe the skill as having 'No database. No storage. No randomness,' implying a purely computational, stateless engine. Later sections explicitly document `add_to_codex()` and say user additions are 'persisted immediately' to `codex.json`, which is a material behavior beyond the claimed no-storage model.

Intent-Code Divergence

Medium
Confidence
96% confidence
Finding
Line L010 states there is 'No storage,' but later documentation says `codex.json` ships with passages and that `add_to_codex()` will compute coordinates and persist them immediately. This is an active contradiction in the skill's own documentation about whether persistent storage exists.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The markdown explicitly says user input can be "computed and persisted immediately" via `add_to_codex()`, which affects local user data by writing to `codex.json`. The file describes the behavior but does not provide a clear user warning or caution about modifying local files or persisting submitted content.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.